Candidate item priority determination method for civil aviation system

Abstract

The invention provides a candidate item priority determination method for a civil aviation system, and relates to the technical field of data processing, and the method comprises the steps: firstly obtaining a basic push score matrix according to a historical data set, and then modifying the basic push score matrix in combination with a score adjustment weight matrix corresponding to a target user, realizing dynamic weight adjustment to obtain a target push score matrix corresponding to the target user; if the first target user data corresponding to the influence factor meets the matching condition corresponding to the influence factor, taking the influence factor as an association factor corresponding to a target user to obtain an association factor set corresponding to the target user; the sum obtained by adding the target push scores of all the correlation factors in the correlation factor set to the candidate items is used as a priority score corresponding to the candidate items; while the global statistical advantage is reserved, the personalized degree and the distinction degree of pushing are remarkably improved, and the distinction degree and the accuracy of the priorities of the candidate items corresponding to different users are enhanced.

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and in particular to a method for determining the priority of candidate items used in a civil aviation system. Background Art

[0002] In the existing civil aviation system, the recommendation order of candidate-related information is usually sorted according to the priority of the candidate, and the information related to the candidate with higher priority is displayed first to improve user interaction efficiency and recommendation effect; in the existing technology, the priority of the candidate can be determined by the following methods: first, statistical analysis of the user behavior data of all users in the historical time period is performed to obtain a fixed priority; second, based on the personalized historical behavior data of the target user, the preference strength of the candidate is predicted through collaborative filtering or deep learning models to obtain personalized priority.

[0003] However, the above method also has the following technical problems: The candidate priority obtained by statistical analysis of the user behavior data of all users in the historical time period is global and can reflect the overall popular trend, but it ignores the individual differences between users and makes it difficult to achieve personalized recommendations; based on the personalized historical behavior data of the target user, the preference intensity of the target user for each candidate is predicted through collaborative filtering or deep learning models, which can generate more personalized recommendation results; however, since the personalized behavior data of the target user is often sparse, the priority generated between different users is relatively small, which affects the discrimination and accuracy of the recommendation results. Summary of the Invention

[0004] In view of the above technical problems, the technical solution adopted by the present invention is: A method for determining the priority of candidate items for a civil aviation system, the method comprising the following steps: S10. Obtain the basic push score matrix A based on the historical data set L; the element a in the i-th row and j-th column of A. ij is the i-th influencing factor X i For the jth candidate H j The basic push score; 1≤i≤m, m is the number of influencing factors; 1≤j≤n, n is the number of candidate items; L includes all data related to each candidate item collected during the historical time period.

[0005] S20. For each influencing factor, if the first target user data corresponding to the influencing factor meets the matching condition corresponding to the influencing factor, the influencing factor is used as the association factor corresponding to the target user Y to obtain the association factor set F corresponding to Y; the first target user data is obtained based on the data acquisition rule corresponding to the matching condition corresponding to the influencing factor.

[0006] S30, according to the second target user data set corresponding to the preset behavior item, the judgment condition corresponding to the preset behavior item and the user label corresponding to Y, obtain the score adjustment weight matrix C corresponding to Y; the element c in the i-th row and j-th column of C ij for a ij The corresponding score adjustment weight is set; the second target user data set includes n second target user data corresponding to the candidate items one by one; the second target user data is obtained based on the data acquisition rule corresponding to the preset behavior item.

[0007] S40. Obtain the target push score matrix D corresponding to Y based on A and C; the element d in the i-th row and j-th column of D ij For X i For H j Target push score; d ij =a ij ×c ij .

[0008] S50: For each candidate item, add the target push scores of all associated factors in F to obtain the sum as the priority score corresponding to the candidate item.

[0009] The present invention has at least the following beneficial effects: The present invention provides a candidate item priority determination method for a civil aviation system. The method can obtain a basic push score matrix based on a historical data set; for each influencing factor, if the first target user data corresponding to the influencing factor meets the matching condition corresponding to the influencing factor, the influencing factor is used as the association factor corresponding to the target user to obtain an association factor set corresponding to the target user; based on the second target user data set corresponding to the preset behavior item, the judgment condition corresponding to the preset behavior item, and the user tag corresponding to the target user, a score adjustment weight matrix corresponding to the target user is obtained; based on the basic push score matrix and the score adjustment weight matrix corresponding to the target user, a target push score matrix corresponding to the target user is obtained, the target push score matrix including the target push score of each influencing factor for each candidate item; for each candidate item, the target push scores of all association factors in the association factor set for the candidate item are added together to obtain a priority score corresponding to the candidate item. It can be seen that the present invention first obtains a basic push score matrix based on a historical data set, and then modifies the basic push score matrix in combination with the score adjustment weight matrix corresponding to the target user, realizing dynamic weight adjustment to obtain a target push score matrix corresponding to the target user; adopts a dynamic weight matrix and a multi-source data fusion mechanism to obtain a target push score matrix; for each influencing factor, if the first target user data corresponding to the influencing factor meets the matching condition corresponding to the influencing factor, then the influencing factor is used as the correlation factor corresponding to the target user to obtain a correlation factor set corresponding to the target user; then the target push scores of all correlation factors in the correlation factor set for the candidate items are added together to obtain the priority score corresponding to the candidate items; while retaining the global statistical advantage, the personalization and discrimination of the push are significantly improved, and the discrimination and accuracy of the priorities of the candidate items corresponding to different users are enhanced, which is further conducive to improving the discrimination and accuracy of the recommendation results. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0011] Figure 1 The present invention provides a flowchart of a method for determining the priority of candidate items in a civil aviation system according to an embodiment of the present invention. DETAILED DESCRIPTION

[0012] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work shall fall within the scope of protection of the present invention.

[0013] It should be noted that the terms "first", "second", etc. in the description and claims of the present invention and the above-mentioned drawings are used to distinguish similar tasks and are not necessarily used to describe a specific order or sequence. It should be understood that the numbers used in this way can be interchanged where appropriate so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product or server that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0014] An embodiment of the present invention provides a method for determining the priority of candidate items for a civil aviation system, the method comprising the following steps: Figure 1 As shown: S10. Obtain the basic push score matrix A based on the historical data set L; the element a in the i-th row and j-th column of A. ij is the i-th influencing factor X i For the jth candidate H j The basic push score; 1≤i≤m, m is the number of influencing factors; 1≤j≤n, n is the number of candidate items; L includes all data related to each candidate item collected during the historical time period; for example: clicks, browsing, collections, purchases, ratings, searches, time of occurrence of the behavior, geographic location, device type, network environment, age, gender, occupation, likes, collections, etc.

[0015] Specifically, the end time point of the historical time period is earlier than the current time point, and the duration of the historical time period is a preset first duration.

[0016] Specifically, candidate items are items that may be recommended or displayed to users, such as products, articles, videos, services, etc.

[0017] Specifically, influencing factors are factors that affect the order in which relevant information of candidate items is pushed, such as hotel preferences, car preferences, bad weather, browsing preferences, air ticket preferences, evening rush hour, and flight itineraries including the top 10 cities for car use.

[0018] Specifically, -10≤a ij≤10.

[0019] Furthermore, a ij The larger the X i For H j The greater the impact.

[0020] Specifically, step S10 includes the following steps S11-S12: S11. Input the historical data set into the g-th preset recommendation model to obtain X i For H j The g-th initial push score W g ij , 1≤g≤h, h is the number of preset recommendation models.

[0021] Specifically, -10≤W g ij ≤10.

[0022] Specifically, the preset recommendation model is a recommendation model predetermined from existing recommendation models. Different preset recommendation models use different recommendation algorithms, for example: a content-based preset recommendation model, a collaborative filtering-based recommendation model, and a rule-based recommendation model.

[0023] S12, based on W g ij Get a ij , a ij Meet the following conditions: a ij =∑ h g=1 (V g ×W g ij ), V g The preset importance weight corresponding to the g-th preset recommendation model.

[0024] Specifically, h g=1 V g =1.

[0025] Through the above steps, the initial push score is obtained based on the historical data set and the preset recommendation model. The basic push score is obtained based on the preset importance weight and initial push score corresponding to the recommendation model. The fusion of multiple models avoids the deviation of a single model and is conducive to improving the accuracy of the obtained basic push score.

[0026] In a specific embodiment, those skilled in the art may obtain the basic push score of each influencing factor for each candidate item by other means, for example, an expert may set the basic push score of each influencing factor for each candidate item based on experience, system prior knowledge or expert knowledge.

[0027] S20. For each influencing factor, if the first target user data corresponding to the influencing factor meets the matching condition corresponding to the influencing factor, the influencing factor is used as the association factor corresponding to the target user Y to obtain the association factor set F corresponding to Y; the first target user data is obtained based on the data acquisition rule corresponding to the matching condition corresponding to the influencing factor.

[0028] Specifically, the first target user data is obtained based on a data acquisition rule corresponding to a matching condition corresponding to an influencing factor, specifically including: For each influencing factor, first target user data corresponding to the influencing factor is obtained from all data related to Y according to a data acquisition rule corresponding to a matching condition corresponding to the influencing factor.

[0029] Specifically, the influencing factors, the matching conditions corresponding to the influencing factors, and the data acquisition rules corresponding to the matching conditions are stored in the first data configuration table corresponding to the influencing factors. The first data configuration table is a configuration table pre-set by technical personnel in this field according to actual needs, including each influencing factor, the matching conditions corresponding to each influencing factor, and the data acquisition rules corresponding to each matching condition; for example: the influencing factor is a severe condition, and the matching condition corresponding to the influencing factor is: the current weather type is any one of heavy snow, freezing, low temperature, strong wind, blowing sand, and heavy rain; the data acquisition rule corresponding to the matching condition is: obtain the weather type of the arrival location of the current trip at the arrival time of the trip.

[0030] In this embodiment, not all influencing factors will affect the priority of the candidate items. For example, if the influencing factor is bad weather, but there is no bad weather at the departure point, arrival point, and transit point of the target user's current itinerary, then the influencing factor will not affect the priority of the candidate items. Through the above steps, the influencing factor will only be used as an association factor when the first target user data meets the matching conditions corresponding to the influencing factor, which can avoid interference from irrelevant factors.

[0031] S30, according to the second target user data set corresponding to the preset behavior item, the judgment condition corresponding to the preset behavior item and the user label corresponding to Y, obtain the score adjustment weight matrix C corresponding to Y; the element c in the i-th row and j-th column of C ij for a ij The corresponding score adjustment weight is set; the second target user data set includes n second target user data corresponding to the candidate items one by one; the second target user data is obtained based on the data acquisition rule corresponding to the preset behavior item.

[0032] Specifically, 0≤c ij ≤10.

[0033] Specifically, the preset behavior items and the judgment conditions corresponding to the preset behavior items are the behavior items and judgment conditions pre-set by technical personnel in this field according to actual needs. For example: the preset behavior item is exposure, and the judgment condition corresponding to the preset behavior item is: no click after ten exposures, no click after five exposures; or the preset behavior items are the first exposure-click ratio and the second exposure-click ratio, and the judgment condition corresponding to the first exposure-click ratio is: the ratio of the number of exposures to the number of clicks is between 0.3-0.5, and the judgment condition corresponding to the second exposure-click ratio is: the ratio of the number of exposures to the number of clicks is between 0.6-0.8.

[0034] Furthermore, the number of preset behavior items is not less than 1.

[0035] Specifically, the number of judgment conditions corresponding to the preset behavior items is not less than 1.

[0036] Specifically, the number of user tags corresponding to Y is not less than 1.

[0037] Specifically, the second target user data is obtained based on the data acquisition rule corresponding to the preset behavior item, which specifically includes the following steps S01-S02: S01. For each preset behavior item, obtain a data acquisition time period corresponding to the preset behavior item, where the end time point of the data acquisition time period is the current time point, and the duration of the data acquisition time period is the preset duration corresponding to the preset behavior item.

[0038] Preferably, the end time point of the historical time period is earlier than the start time point of the data acquisition time period corresponding to any preset behavior item.

[0039] S02, according to the data acquisition rule corresponding to the preset behavior item, collect the data related to Y and H collected during the data acquisition time period corresponding to the preset behavior item j Among all the data that have an associated relationship, obtain the second target user data corresponding to the preset behavior item and the H j The corresponding second target user data.

[0040] Specifically, the data acquisition rules corresponding to the preset behavior items are stored in the second data configuration table corresponding to the preset behavior items. The second data configuration table is a configuration table pre-set by technical personnel in this field according to actual needs, including each preset behavior item and the data acquisition rules corresponding to each preset behavior item.

[0041] Through the above steps, the data acquisition time period corresponding to each preset behavior item is obtained, and the second target user data corresponding to the preset behavior item is obtained based on the data acquisition time period and the data acquisition rules corresponding to the preset behavior item. This can filter out irrelevant or invalid data, avoid processing irrelevant or invalid data, and waste resources.

[0042] Specifically, step S30 includes the following sub-steps S31-S33: S31, for each preset behavior item, if the second target user data corresponding to the preset behavior item is concentrated and H j The corresponding second target user data meets the judgment condition corresponding to the preset behavior item, then according to the weight update rule corresponding to the judgment condition, each influencing factor corresponding to the preset behavior item is updated. j The basic adjustment weight corresponding to the basic push score is updated to update E; E is the basic adjustment matrix, and the element on the i-th row and j-th column in E is a ij The corresponding basic adjustment weight.

[0043] Specifically, the weight update rules corresponding to the judgment conditions are update rules pre-set by technical personnel in this field according to actual needs, for example: multiply by 0.2; multiply by 6; multiply by 3; multiply by 0.5; add 2; add 5, subtract 2; no further details will be given here.

[0044] Specifically, each update of the basic adjustment weight is based on the latest basic adjustment weight obtained from the previous update. According to the priority or execution order of the preset behavior items, the basic adjustment weights corresponding to the basic push scores of the candidate items of each influencing factor corresponding to the preset behavior item are iteratively adjusted in turn to ensure that the update process is timely and continuous; the priority or execution order of the preset behavior items are predetermined by technical personnel in this field according to actual needs and will not be repeated here.

[0045] Specifically, the initial value of the basic adjustment weight is 1.

[0046] Specifically, the corresponding relationship between the preset behavior items and the influencing factors is predetermined by those skilled in the art according to actual needs and will not be elaborated here.

[0047] S32, after all updates, E is used as the first adjustment matrix G corresponding to M; the element in the i-th row and j-th column of G is a ij The corresponding first adjustment weight.

[0048] S33. Update G according to the user tag corresponding to Y to obtain C.

[0049] Through the above steps, according to the preset behavior items and the second target user data corresponding to the preset behavior items, the basic adjustment matrix is ​​updated and iterated multiple times to obtain a first adjustment weight matrix, and the first adjustment weight matrix is ​​further updated according to the user label corresponding to the target user to obtain a score adjustment weight matrix corresponding to the target user; the score adjustment weight matrix is ​​determined according to the characteristics of the target user, which is conducive to improving the personalization and discrimination of the obtained score adjustment weight matrix.

[0050] Specifically, step S33 includes the following steps S331-S332: S331. For each user tag corresponding to Y, if the user tag is the same as any preset tag, then according to the weight update rule corresponding to the preset tag that is the same as the user tag, the first adjustment weight corresponding to the basic push score of each candidate item for each influencing factor corresponding to the user tag is updated, so that G is updated.

[0051] Specifically, the weight update rules corresponding to the preset labels are update rules pre-set by technical personnel in this field according to actual needs, for example: multiply by 0.2; multiply by 6; multiply by 3; multiply by 0.5; add 2; add 5, subtract 3; set to 4; set to 0; no further details will be given here.

[0052] Specifically, each update of the first adjustment weight is based on the latest first adjustment weight obtained from the previous update. According to the priority or execution order of the user tag, the first adjustment weight corresponding to the basic push score of the candidate item of each influencing factor corresponding to the user tag is iteratively adjusted in turn to ensure that the update process is timely and continuous; the priority or execution order of the user tag is predetermined by technical personnel in this field according to actual needs and will not be repeated here.

[0053] Specifically, the influencing factors corresponding to the user tag are consistent with the influencing factors corresponding to the preset tag that is the same as the user tag.

[0054] Specifically, the correspondence between the preset labels and the influencing factors is pre-set by those skilled in the art according to actual needs, and will not be described in detail here.

[0055] S332. After all updates are completed, G is used as C.

[0056] Through the above steps, the first adjustment weight matrix is ​​updated and iterated multiple times according to the user label of the target user to obtain the score adjustment weight matrix corresponding to the target user, which is conducive to improving the personalization and discrimination of the obtained score adjustment weight matrix.

[0057] S40. Obtain the target push score matrix D corresponding to Y based on A and C; the element d in the i-th row and j-th column of D ij For X i For H j Target push score; d ij =a ij ×c ij .

[0058] S50: For each candidate item, add the target push scores of all associated factors in F to obtain a sum as the priority score corresponding to the candidate item.

[0059] Specifically, the greater the priority score of a candidate item, the higher the candidate item priority corresponding to the candidate item.

[0060] Furthermore, the candidate item priority is used to determine the order in which the relevant information of the candidate items is pushed, wherein the relevant information of the candidate items with a high candidate item priority is pushed to M first.

[0061] Through the above steps, first, a basic push score matrix is ​​obtained based on the historical data set; for each influencing factor, if the first target user data corresponding to the influencing factor meets the matching condition corresponding to the influencing factor, the influencing factor is used as the correlation factor corresponding to the target user to obtain the correlation factor set corresponding to the target user; the influence of irrelevant influencing factors is avoided; based on the second target user data set corresponding to the preset behavior item, the judgment condition corresponding to the preset behavior item and the user label corresponding to the target user, the score adjustment weight matrix corresponding to the target user is obtained; based on the basic push score matrix and the score corresponding to the target user, the target push score matrix corresponding to the target user is obtained. The target push score matrix includes the target push score of each influencing factor for each candidate item; the basic push score matrix is ​​modified in combination with the score adjustment weight matrix corresponding to the target user, and the dynamic weight matrix and multi-source data fusion mechanism are used to obtain the target push score matrix corresponding to the target user; the sum of the target push scores of all associated factors in the associated factor set for the candidate item is added as the priority score corresponding to the candidate item; while retaining the global statistical advantage, the personalization and discrimination of the push are significantly improved, and the discrimination and accuracy of the priority of the candidate items corresponding to different users are enhanced, which is further conducive to improving the discrimination and accuracy of the recommendation results.

[0062] In a specific embodiment, the following steps are further included after step S32: S321: Determine, based on a data validity judgment condition corresponding to a preset behavior item, whether each second target user data corresponding to the preset behavior item is valid data.

[0063] Specifically, the data validity judgment condition corresponding to the preset behavior item is a condition pre-set by technical personnel in this field according to actual needs for judging whether the data is valid. For example: the preset behavior item is exposure, and the data validity judgment corresponding to the preset behavior item is that the exposure duration is greater than the preset minimum exposure duration; if the data in the second target user data corresponding to the preset behavior item indicates that the exposure duration is not greater than the preset minimum exposure duration, then the second target user data is determined to be invalid data; if the data in the second target user data corresponding to the preset behavior item indicates that the exposure duration is greater than the preset minimum exposure duration, then the second target user data is determined to be valid data.

[0064] S322, for Hj , if the second target user data corresponding to all preset behavior items are concentrated, and H j The corresponding second target user data are all invalid data, then determine H j Data label H_G in G j is invalid; otherwise, determine H_G j is valid.

[0065] S323, according to G, H_G j Get the user tag C corresponding to Y.

[0066] In this embodiment, when the target user is a new user, or the target user unintentionally clicks or swipes over certain content, some invalid data may be generated; through the above steps, whether the second target user data is valid data is determined according to the data validity judgment condition corresponding to the preset behavior item. If the second target user data corresponding to all the preset behavior items are invalid data, then the first adjustment weight corresponding to the basic push score of the candidate item by each influencing factor obtained based on these invalid second target user data is also invalid. Therefore, the data label of the candidate item in the first adjustment weight matrix is ​​determined to be invalid. Otherwise, the data label of the candidate item in the first adjustment weight matrix is ​​determined to be valid. The score adjustment weight matrix corresponding to the target user is obtained according to the data label of the candidate item in the first adjustment weight matrix, which is conducive to improving the accuracy of the score adjustment weight matrix.

[0067] Specifically, S323 also includes the following steps: S100, when there are at least α candidates whose data labels in G are valid and not all candidates whose data labels in G are valid, if H_G j If it is invalid, then determine H j is the first candidate; if H_G j If it is valid, then determine H j is the second candidate; to obtain the first candidate list L=(L1, L2, ..., L k ,…,L t ) and the second candidate list P=(P1, P2, ..., P y ,…,P q ); α is the number of preset valid items; L k is the kth first candidate, 1≤k≤t, t is the number of first candidates, P y is the yth second candidate, 1≤y≤q, and q is the number of second candidates.

[0068] Specifically, 1<α<n.

[0069] Specifically, t+q=n.

[0070] Specifically, step S100 further includes: if the data labels of each candidate item in G are all valid or invalid, then G is used as R.

[0071] S200: Select L from the first adjustment matrices corresponding to all other users obtained within a preset time period. k The corresponding first intermediate matrix list Q k =(Q k1 , Q k2 ,…,Q kr ,…Q ks(k) ), Q kr For L k The corresponding r-th first intermediate matrix, 1≤r≤s(k), s(k) is L k The number of corresponding first intermediate matrices; P1, P2, ..., P y ,…,P q and L k In Q kr The data labels in are all valid; it can be understood as: Q kr Belongs to a set constructed by first adjustment matrices corresponding to all other users acquired within a preset time period.

[0072] Specifically, the other user is any user except the target user.

[0073] Specifically, the end time point of the preset time period is the current time point, the duration of the preset time period is the preset second duration, and the preset second duration is smaller than the preset first duration; preferably, the end time point of the preset time period is the current time point.

[0074] S300, if the second intermediate matrix G corresponding to G 0 With Q kr The corresponding third intermediate matrix Q 0 kr The similarity between them is greater than the preset similarity threshold S 0 , then Q kr As L k The corresponding first key matrix to obtain L k The corresponding first key matrix list V k =(V k1 , V k2 ,…,V kz ,…,V ku(k) ), V kz For L k The corresponding z-th first key matrix, 1≤z≤u(k), u(k) is L k The number of the corresponding first key matrix; G 0 and Q0 kr are all m×q matrices and G 0 The element in row i and column y is X in G i P y The first adjustment weight corresponding to the basic push score, Q 0 kr The element in row i and column y is Q kr Medium X i P y A first adjustment weight corresponding to the basic push score; Those skilled in the art know that any method of obtaining the similarity between two matrices in the prior art falls within the scope of protection of the present invention and will not be described in detail here.

[0075] Specifically, G 0 With Q 0 kr The greater the similarity between G 0 With Q 0 kr The more similar.

[0076] Specifically, 0.8≤S 0 ≤1.

[0077] S400, V k1 , V k2 ,…,V kz ,…,V ku(k) Medium X i To L k The weighted average of the first adjustment weights corresponding to the basic push scores is used as the X i To L k The first adjustment weight corresponding to the basic push score is used to update G and use G after all updates as the second adjustment matrix R corresponding to M.

[0078] S500: Update R according to the user tag corresponding to Y to obtain C.

[0079] Through the above steps, when there are at least α candidates whose data labels in the first adjustment weight matrix corresponding to the target user are valid and not all candidates have valid data labels in the first adjustment weight matrix, the first adjustment weight of the target user can be updated according to the first adjustment weights corresponding to other users, and the candidate whose data label in the first adjustment weight matrix corresponding to the target user is invalid is used as the first candidate, and the candidate whose data label in the second adjustment weight matrix corresponding to the target user is valid is used as the second candidate. For each first candidate, the first intermediate candidate whose data labels are valid is screened out from the first adjustment matrix corresponding to other users. According to the similarity between the second intermediate matrix and the third intermediate matrix, a first key matrix is ​​screened out, in which the data label of the first candidate item is valid and relatively similar to the first adjustment matrix corresponding to the target user; based on the first adjustment weight corresponding to the basic push score of the candidate item by each influencing factor in the first key matrix, the first adjustment weight corresponding to the basic push score of the candidate item by each influencing factor in the first adjustment weight matrix corresponding to the target user is iteratively updated; and the first adjustment weight matrix after all updates is used as the second adjustment weight matrix corresponding to the target user; a compensation mechanism is adopted to improve the accuracy of the second adjustment weight matrix, and further, the accuracy of the score adjustment weight matrix is ​​improved.

[0080] Specifically, step S500 also includes the following steps: S501. For each user tag corresponding to Y, if the user tag is the same as any preset tag, then according to the weight update rule corresponding to the preset tag that is the same as the user tag, the second adjustment weight corresponding to the basic push score of each candidate item for each influencing factor corresponding to the user tag is updated, so that R is updated.

[0081] Specifically, each update of the second adjustment weight is based on the latest second adjustment weight obtained from the previous update. According to the priority or execution order of the user tags, the second adjustment weight corresponding to the basic push score of the candidate item of each influencing factor corresponding to the user tag is iteratively adjusted in turn to ensure that the update process is timely and continuous.

[0082] S502: Use R after all updates as C.

[0083] Through the above steps, the second adjustment weight matrix is ​​updated and iterated multiple times according to the user label of the target user to obtain the score adjustment weight matrix corresponding to the target user, which is conducive to improving the personalization and discrimination of the obtained score adjustment weight matrix.

[0084] In a specific embodiment, after step S200 and before step S500, the following steps are further included to obtain R: S210, if the first specified matrix G corresponding to G 1 With Q kr The corresponding second specified matrix Q 1 kr The similarity between them is greater than S 0 , then Q kr As L k The corresponding second key matrix to obtain L k The corresponding second key matrix list W k =(W k1 , W k2 ,…,W kβ ,…,W kθ(k) ), W kβ For L k The corresponding β-th second key matrix, 1≤β≤θ(k), θ(k) is L k The number of the corresponding second key matrix; G 1 and Q 1 kr are all matrices of f×q and G 1 The element in row b and column y is F in G b P y The first adjustment weight corresponding to the basic push score, Q 1 kr The element in row b and column y is Q kr Medium F b P y The first adjustment weight corresponding to the basic push score is f, f is the number of related factors in F, F b is the b-th correlation factor in F, 1≤b≤f.

[0085] Specifically, G 1 With Q 1 kr The greater the similarity between G 1 With Q 1 kr The more similar.

[0086] S220, W k1 , W k2 ,…,W kβ ,…,W kθ(k) Medium F b To L k The weighted average of the first adjustment weight corresponding to the basic push score is used as the F in G b To L kThe first adjustment weight corresponding to the basic push score is used to update G and use G after all updates as R.

[0087] Through the above steps, since not all influencing factors will affect the priority of the candidate items, according to the similarity between the first specified matrix and the second specified matrix, the second key matrix whose data label of the first candidate item is valid and relatively similar to the first adjustment matrix corresponding to the target user can be screened out, and based on the first adjustment weight corresponding to the basic push score of the candidate item by each correlation factor in the second key matrix, the first adjustment weight corresponding to the basic push score of the candidate item by each correlation factor in the first adjustment weight matrix corresponding to the target user is iteratively updated; and the first adjustment weight matrix after all updates is completed is used as the second adjustment weight matrix corresponding to the target user; compared with the above embodiment, the amount of data to be processed is reduced, which is conducive to improving calculation efficiency, and at the same time, it can also reduce the influence of irrelevant influencing factors, which is conducive to improving the accuracy of the second adjustment weight matrix, and further, it is conducive to improving the accuracy of the score adjustment weight matrix.

[0088] In a specific embodiment, the following steps are further included after step S50: S60 . For each of the plurality of candidate adjustment rules, obtain a data set corresponding to the candidate adjustment rule, where the data set includes a question text, at least one influencing factor, at least one candidate item, and a score adjustment method.

[0089] Specifically, the candidate adjustment rules are predetermined by those skilled in the art based on actual needs, for example, vehicle service enhancement recommendation rules.

[0090] Specifically, the question text includes: inquiry text; the inquiry text is a text pre-set by technical personnel in this field based on the candidate adjustment rules to clarify the decision-making goals, for example: the candidate adjustment rule is the car service enhancement recommendation rule, and the inquiry included in the question text corresponding to the candidate adjustment rule is: whether to adjust the priority score of the candidate items related to the car service.

[0091] Specifically, the question text also includes prompt text, which is used to guide the large language model to obtain information related to the decision-making target corresponding to the inquiry text corresponding to the prompt text, and help the large language model understand which external data should be collected for auxiliary judgment. For example: Please comprehensively judge whether it is necessary to enhance the priority score of candidates related to car services based on factors such as the current user's geographic location, local car use policies, whether there are major events or holidays, etc.

[0092] Specifically, after receiving the question text, the large language model automatically calls external data sources based on the question text, such as weather API, map API, local database, etc., integrates and infers information, and outputs the decision result corresponding to the question text.

[0093] Furthermore, the correspondence between the candidate adjustment rules and the influencing factors, as well as the correspondence between the candidate adjustment rules and the candidate items are predetermined by technical personnel in this field based on actual needs. For example, the candidate adjustment rule is a car service enhancement recommendation rule, and the influencing factors corresponding to the candidate adjustment rule are flight itineraries including the top 10 car-using cities and evening rush hour; the candidate items corresponding to the candidate adjustment rule are vehicle reservation services, real-time traffic information push services, and carpooling services.

[0094] Furthermore, the score adjustment method corresponding to the candidate adjustment rule is an adjustment method pre-set by technical personnel in this field based on the candidate adjustment rule. For example: the candidate adjustment rule is a car service enhancement recommendation rule, and the score adjustment method corresponding to the candidate adjustment rule is: adjusting the priority score to 100; adjusting the priority score to 90; no further details will be given here.

[0095] S70 . For each candidate adjustment rule, if F includes all influencing factors corresponding to the candidate adjustment rule, then the question text corresponding to the candidate adjustment rule is input into the large language model to obtain a decision result J corresponding to the question text.

[0096] S80. If J is a first-category result, the priority scores corresponding to all candidate items corresponding to the candidate adjustment rule are adjusted according to the score adjustment method corresponding to the candidate adjustment rule corresponding to J to obtain the target priority score corresponding to each candidate item; the first-category result indicates that the priority of the candidate items corresponding to the candidate adjustment rule needs to be adjusted.

[0097] Specifically, the decision result also includes a second type of result, and the second type of result indicates that there is no need to adjust the priority of the candidate items corresponding to the candidate adjustment rule.

[0098] Specifically, the greater the target priority score of a candidate item, the higher the candidate item priority corresponding to the candidate item.

[0099] Through the above steps, compared with manually adjusting the priority scores of candidate items according to actual needs, the embodiment of the present application realizes the adjustment of priority scores through a large language model. In the embodiment of the present application, a data set corresponding to the candidate adjustment rule is obtained, and the data set includes a question text, at least one influencing factor, at least one candidate item and a score adjustment method. If the associated factor set includes all the influencing factors corresponding to the candidate adjustment rule, the question text corresponding to the candidate adjustment rule is input into the large language model to obtain the decision result corresponding to the question text; if the decision result is a first-class result, the priority scores corresponding to all candidate items corresponding to the candidate adjustment rule are adjusted according to the score adjustment method corresponding to the candidate adjustment rule corresponding to the decision result to obtain the target priority score corresponding to each candidate item; the automated and intelligent adjustment of the candidate item priority scores is realized, which not only improves the accuracy of the target priority score, but also effectively reduces the manual maintenance cost and significantly improves the data processing efficiency.

[0100] An embodiment of the present invention also provides a non-transitory computer-readable storage medium, which can be set in an electronic device to store a computer program related to a method in the method embodiment. The computer program is loaded and executed by the processor to implement the method provided in the above embodiment.

[0101] An embodiment of the present invention further provides an electronic device, comprising: a processor, a memory, and a computer program stored in the memory and executable on the processor, wherein the processor implements the method provided in the above embodiment when executing the computer program.

[0102] An embodiment of the present invention further provides a computer program product comprising program code. When the program product is run on an electronic device, the program code is used to enable the electronic device to execute the steps of the method according to various exemplary embodiments of the present invention described above in this specification.

[0103] The present invention provides a candidate item priority determination method for a civil aviation system. The method can obtain a basic push score matrix based on a historical data set; for each influencing factor, if the first target user data corresponding to the influencing factor meets the matching condition corresponding to the influencing factor, the influencing factor is used as the association factor corresponding to the target user to obtain an association factor set corresponding to the target user; based on the second target user data set corresponding to the preset behavior item, the judgment condition corresponding to the preset behavior item, and the user tag corresponding to the target user, a score adjustment weight matrix corresponding to the target user is obtained; based on the basic push score matrix and the score adjustment weight matrix corresponding to the target user, a target push score matrix corresponding to the target user is obtained, the target push score matrix including the target push score of each influencing factor for each candidate item; for each candidate item, the target push scores of all association factors in the association factor set for the candidate item are added together to obtain a priority score corresponding to the candidate item. It can be seen that the present invention first obtains a basic push score matrix based on a historical data set, and then modifies the basic push score matrix in combination with the score adjustment weight matrix corresponding to the target user, realizing dynamic weight adjustment to obtain a target push score matrix corresponding to the target user; adopts a dynamic weight matrix and a multi-source data fusion mechanism to obtain a target push score matrix; for each influencing factor, if the first target user data corresponding to the influencing factor meets the matching condition corresponding to the influencing factor, then the influencing factor is used as the correlation factor corresponding to the target user to obtain a correlation factor set corresponding to the target user; then the target push scores of all correlation factors in the correlation factor set for the candidate items are added together to obtain the priority score corresponding to the candidate items; while retaining the global statistical advantage, the personalization and discrimination of the push are significantly improved, and the discrimination and accuracy of the priorities of the candidate items corresponding to different users are enhanced, which is further conducive to improving the discrimination and accuracy of the recommendation results.

[0104] Although some specific embodiments of the present invention have been described in detail by way of examples, it should be understood by those skilled in the art that the above examples are for illustration only and are not intended to limit the scope of the present invention. It should also be understood by those skilled in the art that various modifications may be made to the embodiments without departing from the scope and spirit of the present invention.

Claims

1. A method for determining the priority of candidate items for a civil aviation system, characterized in that: The method comprises the following steps: S10. Obtain the basic push score matrix A based on the historical data set L; the element a in the i-th row and j-th column of A. ij is the i-th influencing factor X i For the jth candidate H j Basic push score; 1≤i≤m, m is the number of influencing factors; 1≤j≤n, n is the number of candidate items; L includes all data related to each candidate item collected during the historical time period; S20. For each influencing factor, if the first target user data corresponding to the influencing factor satisfies the matching condition corresponding to the influencing factor, the influencing factor is used as the association factor corresponding to the target user Y to obtain an association factor set F corresponding to Y; the first target user data is obtained based on the data acquisition rule corresponding to the matching condition corresponding to the influencing factor; S30, according to the second target user data set corresponding to the preset behavior item, the judgment condition corresponding to the preset behavior item and the user label corresponding to Y, obtain the score adjustment weight matrix C corresponding to Y; the element c in the i-th row and j-th column of C ij for a ij The corresponding score adjustment weight; The second target user data set includes n pieces of second target user data corresponding one-to-one to the candidate items; The second target user data is obtained based on the data acquisition rule corresponding to the preset behavior item; S40. Obtain the target push score matrix D corresponding to Y based on A and C; the element d in the i-th row and j-th column of D ij For X i For H j Target push score; d ij =a ij ×c ij ; S50: For each candidate item, add the target push scores of all associated factors in F to obtain a sum as the priority score corresponding to the candidate item.

2. The candidate item priority determination method for a civil aviation system according to claim 1, characterized in that: The first target user data is obtained based on a data acquisition rule corresponding to a matching condition corresponding to an influencing factor, specifically including: For each influencing factor, first target user data corresponding to the influencing factor is obtained from all data related to Y according to a data acquisition rule corresponding to a matching condition corresponding to the influencing factor.

3. The candidate item priority determination method for a civil aviation system according to claim 1, characterized in that: The second target user data is obtained based on the data acquisition rule corresponding to the preset behavior item, specifically including the following steps: S01. For each preset behavior item, obtain a data acquisition time period corresponding to the preset behavior item, where the end time point of the data acquisition time period is the current time point, and the duration of the data acquisition time period is the preset duration corresponding to the preset behavior item; S02, according to the data acquisition rule corresponding to the preset behavior item, collect the data related to Y and H collected during the data acquisition time period corresponding to the preset behavior item j Among all the data that have an associated relationship, obtain the second target user data corresponding to the preset behavior item and the H j The corresponding second target user data.

4. The candidate item priority determination method for a civil aviation system according to claim 1, characterized in that: Step S30 includes the following sub-steps: S31, for each preset behavior item, if the second target user data corresponding to the preset behavior item is concentrated and H j The corresponding second target user data meets the judgment condition corresponding to the preset behavior item, then according to the weight update rule corresponding to the judgment condition, each influencing factor corresponding to the preset behavior item is updated. j The basic adjustment weight corresponding to the basic push score is updated to update E; E is the basic adjustment matrix, and the element on the i-th row and j-th column in E is a ij The corresponding basic adjustment weight; S32, after all updates, E is used as the first adjustment matrix G corresponding to M; the element in the i-th row and j-th column of G is a ij the corresponding first adjustment weight; S33. Update G according to the user tag corresponding to Y to obtain C.

5. The candidate item priority determination method for a civil aviation system according to claim 4, characterized in that: Step S33 includes the following steps: S331. For each user tag corresponding to Y, if the user tag is the same as any preset tag, then according to the weight update rule corresponding to the preset tag that is the same as the user tag, update the first adjustment weight corresponding to the basic push score of each candidate item for each influencing factor corresponding to the user tag, so that G is updated; S332. After all updates are completed, G is used as C.

6. The candidate item priority determination method for a civil aviation system according to claim 1, characterized in that: The larger the priority score of a candidate item, the higher the candidate item priority corresponding to the candidate item.

7. The candidate item priority determination method for a civil aviation system according to claim 6, characterized in that: The candidate item priority is used to determine the order in which the relevant information of the candidate items is pushed, wherein the relevant information of the candidate items with a high candidate item priority is pushed to M first.

8. The candidate item priority determination method for a civil aviation system according to claim 1, characterized in that: -10≤a ij ≤10。 9. The method for determining candidate priority for a civil aviation system according to claim 1, wherein: 0≤c ij ≤10。 10. The candidate item priority determination method for a civil aviation system according to claim 4, characterized in that: The initial value of the basic adjustment weight is 1.

Images